Data processing system, system-build system, and system-build method

• US 5,983,210 A
• Filed: 12/23/1996
• Issued: 11/09/1999
• Est. Priority Date: 12/27/1995
• Status: Expired due to Fees

• Current Assignee: Kabushiki Kaishi Toshiba
• Original Assignee: Kabushiki Kaisha Toshiba (Toshiba Corporation)
• Inventors: Aisu, Hideyuki, Sekine, Satoshi, Makino, Kyoko, Kano, Makoto, Imasaki, Naoki
• Primary Examiner(s): Downs, Robert W.
• Assistant Examiner(s): Rhodes, Jason W.
• Application Number:
  US08/772,297
• Time in Patent Office: 1,051 Days
• Field of Search: 395/21, 395/22, 395/23, 395/24, 395/27, 395/11, 706/15-17, 706/25-27, 706/41, 706/43, 706/10, 706/44
• US Class Current: 706/15
• CPC Class Codes: G06N 3/08 Learning methods

Abstract

Each of a plurality of units 2 has the learning section 9 which updates the parameters PRM and performs learning. The learning means 9 modifies the parameters PRM based on the error signal OS'"'"' propagated from one unit to another. To build a system consisting of such units 2, the user uses the GUI or some other method to select the units 2 as well as the signal paths for connecting the units via the user interface, displays the system configuration status, and adjusts the functions of units 2. Therefore, an easy-to-build data processing system and a system-build system and method used to build such a system can be provided.

Claims

1. A data processing system comprising:

a plurality of data processing units coupled together to form a first layered neural network, the data processing units generating output signals by modifying an input signal according to parameters associated with each of the data processing units and transmitting the output signal in an upstream direction through the layered network, the data processing units transmitting error signals in a downstream direction through the layered network, each said data processing unit further including

modification means for modifying said parameters of the data processing unit based on the received error signal, the received error signals being based on a difference between a final output signal of the network and a desired output signal, and

at least one of the data processing units further comprising

a second plurality of data processing units coupled together to form a second layered neural network, wherein the first and second layered networks form a hierarchy of networks in which the second layered network is subordinate to the first layered network, and wherein the first and second layered networks each further include a unit control means for controlling an operating sequence of data processing units within each layered network.

2. The data processing system of claim 1, wherein the modification means further comprises means for generating a new error signal based on the received error signal.

3. The data processing system of claim 1, wherein the data processing units each further comprise:

control means for controlling the operation of the data processing unit, the control means of data processing units in a subordinate network controlling the operation of the data processing units in the subordinate network based on an operation instruction transmitted from the unit control means of a superordinate network.